本文選題：視頻質(zhì)量評價 + 全參考　； 參考：《天津大學(xué)》2016年碩士論文

【摘要】：由于網(wǎng)絡(luò)視頻應(yīng)用的快速發(fā)展,數(shù)字視頻逐漸走近我們的日常生活,如視頻點播、數(shù)字電視、視頻會議、網(wǎng)絡(luò)流媒體視頻和視頻監(jiān)控等。從前端視頻獲取設(shè)備到用戶終端,為了保證視頻應(yīng)用服務(wù)質(zhì)量和用戶體驗,我們需要從視頻處理和傳輸?shù)拳h(huán)節(jié)檢測視頻圖像質(zhì)量。由于視頻影像最終受體往往是人眼,任何一個有生活經(jīng)驗的普通人都能較為準(zhǔn)確地評價視頻的清晰度、連貫性、顏色的鮮艷程度、圖像飽和度等。因此人眼主觀評價是最準(zhǔn)確的視頻質(zhì)量評價途徑。然而,對于如今巨大的視頻數(shù)據(jù)量,主觀視頻質(zhì)量評價已經(jīng)力不從心。因此采用數(shù)學(xué)方法和計算機程序進行自動化視頻質(zhì)量評價的客觀評價方法,已經(jīng)成為視頻質(zhì)量評價領(lǐng)域的研究熱點�？陀^視頻質(zhì)量評價方法根據(jù)算法對原始視頻的依賴程度分為全參考、部分參考和無參考評價算法。本文研究重點是全參考與無參考視頻質(zhì)量評價算法,主要工作分為三個部分:1.在對視頻空域與時域視覺感知特性研究的基礎(chǔ)上,引入三維梯度相似度改進了全參考視頻質(zhì)量評價算法——基于時空域梯度相似度的視頻質(zhì)量評價算法,并將其與當(dāng)前國際上的全參考視頻質(zhì)量評價算法進行比較,實驗結(jié)果表明本算法具有較好的評價性能和較低算法復(fù)雜度;2.對國際上當(dāng)前最好的通用型無參考圖像質(zhì)量評價算法進行綜述。首先介紹了每一種算法的特征提取和質(zhì)量評價原理,然后在LIVE主觀質(zhì)量評價數(shù)據(jù)庫上對幾種算法進行了仿真實驗,最后定量地和定性地評價了這幾種算法的性能,分析了各自算法的優(yōu)勢與不足;3.調(diào)研了視頻監(jiān)控系統(tǒng)中存在的常見視頻空域、時域失真類型,并將通用型無參考圖像質(zhì)量評價算法應(yīng)用到視頻監(jiān)控系統(tǒng)中,最后將客觀評價結(jié)果與主觀評價結(jié)果進行了對比。
[Abstract]:With the rapid development of network video applications, digital video gradually approaches our daily life, such as video-on-demand, digital television, video conferencing, network streaming video and video surveillance. From the front-end video acquisition device to the user terminal, in order to ensure the quality of service and user experience of video application, we need to detect the video image quality from video processing and transmission. Because the ultimate receptor of video image is usually human eye, any ordinary person with life experience can accurately evaluate the definition, consistency, bright color, image saturation and so on. Therefore, subjective evaluation of human eyes is the most accurate way to evaluate video quality. However, subjective video quality evaluation has been inadequate for the huge amount of video data. Therefore, the objective evaluation method of automatic video quality evaluation using mathematical methods and computer programs has become a research hotspot in the field of video quality evaluation. Objective video quality evaluation method is divided into full reference, partial reference and non-reference evaluation algorithm according to the degree of dependence of the algorithm on the original video. This paper focuses on the full reference and no reference video quality evaluation algorithm, the main work is divided into three parts: 1. Based on the research of visual perception characteristics in spatial domain and time domain, 3D gradient similarity is introduced to improve the full reference video quality evaluation algorithm, which is based on temporal and spatial gradient similarity. The experimental results show that this algorithm has better evaluation performance and lower complexity. This paper reviews the best universal image quality evaluation algorithms in the world. Firstly, the principle of feature extraction and quality evaluation of each algorithm is introduced, then several algorithms are simulated on the LIVE subjective quality evaluation database. Finally, the performance of these algorithms is evaluated quantitatively and qualitatively. The advantages and disadvantages of each algorithm are analyzed. The common video spatial and temporal distortion types in video surveillance system are investigated, and the general non-reference image quality evaluation algorithm is applied to the video surveillance system. Finally, the objective evaluation results are compared with the subjective evaluation results.
【學(xué)位授予單位】：天津大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2016
【分類號】：TP391.41

【參考文獻】

相關(guān)期刊論文 前9條

1 王秋;胡彥;陳秋華;;保證高清視頻監(jiān)控系統(tǒng)質(zhì)量的要點分析[J];中國安防;2014年11期

2 桑慶兵;蘇媛媛;李朝鋒;吳小俊;;基于梯度結(jié)構(gòu)相似度的無參考模糊圖像質(zhì)量評價[J];光電子.激光;2013年03期

3 邵宇;孫富春;劉瑩;;基于局部結(jié)構(gòu)張量的無參考型圖像質(zhì)量評價方法[J];電子與信息學(xué)報;2012年08期

4 周武杰;郁梅;蔣剛毅;彭宗舉;邵楓;;基于視覺感知和零水印的部分參考立體圖像質(zhì)量評價模型[J];電子與信息學(xué)報;2012年08期

5 范媛媛;沈湘衡;桑英軍;;基于對比度敏感度的無參考圖像清晰度評價[J];光學(xué)精密工程;2011年10期

6 路文;高新波;王體勝;;一種基于小波分析的部分參考型圖像質(zhì)量評價方法[J];電子與信息學(xué)報;2009年02期

7 葉盛楠;蘇開娜;肖創(chuàng)柏;段娟;;基于結(jié)構(gòu)信息提取的圖像質(zhì)量評價[J];電子學(xué)報;2008年05期

8 王體勝;高新波;路文;李廣東;;一種新的部分參考型圖像質(zhì)量評價方法[J];西安電子科技大學(xué)學(xué)報;2008年01期

9 楊春玲;陳冠豪;謝勝利;;基于梯度信息的圖像質(zhì)量評判方法的研究[J];電子學(xué)報;2007年07期

，

本文編號：1856451